Accurate classification and understanding of metabolite variations in Eurycoma longifolia according to source, harvesting age, and planting location are of great importance to ensure the quality and standardization of E. longifolia-based products. In this study, an untargeted metabolomic approach using high-performance liquid chromatography coupled to quadrupole-time-of-flight mass spectrometry (HPLC-Q-ToF-MS) was used to classify E. longifolia samples and to determine metabolite variations in 1-, 4-, and 11-year-old E. longifolia roots collected from different planting locations. Principal component analysis (PCA) and hierarchical clustering analysis showed clear separation between hairy root culture and E. longifolia root samples, highlighting apparent metabolite differences between these samples. Moreover, the approach successfully separated E. longifolia root samples according to their harvesting age and planting location. The PCA loading plot after varimax rotation revealed that dominant ions with strong loadings to principal component 1 were identified as canthin-6-one alkaloid and quassinoid compounds. Overall, the findings indicate that HPLC-Q-ToF-MS-based metabolomics is a powerful approach to classify and determine metabolite variations in E. longifolia samples.